ACFM3506.pdf (19.72Mb)

(1)

ICES A

DVISORY

C

OMMITTEE ON

F

ISHERY

M

ANAGEMENT

ACFM:35

R EPORT OF THE W ^ORKING G ROUP ON THE

A SSESSMENT OF D ^EMERSAL S TOCKS IN THE

N ^ORTH S ^{EA AND} S ^KAGERRAK (WGNSSK)

5–14 S EPTEMBER 2006

ICES H EADQUARTERS

(2)

DK-1553 Copenhagen V Denmark

Telephone (+45) 33 38 67 00 Telefax (+45) 33 93 42 15 www.ices.dk

[email protected]

Recommended format for purposes of citation:

ICES. 2006. Report of the Working Group on the Assessment of Demersal Stocks in the North Sea and Skagerrak (WGNSSK), 5–14 September 2006, ICES Headquarters. ACFM:35. 1160 pp.

For permission to reproduce material from this publication, please apply to the General Secretary.

The document is a report of an Expert Group under the auspices of the International Council for the Exploration of the Sea and does not necessarily represent the views of the Council.

(3)

The ICES Working Group for the Assessment of Demersal Stocks in the North Sea and Skagerrak (WGNSSK) met at ICES Headquarters in Copenhagen, Denmark, during 5-14 September 2006. There were 30 participants from 8 countries. The main terms of reference for the Working Group were: to carry out stock assessments and to provide catch forecasts for demersal and industrial stocks in the North Sea, Skagerrak and Eastern Channel; to consider environmental drivers of fish population dynamics and effects of fisheries on ecosystems; to collate data for mixed fisheries evaluations; to evaluate stock recovery and management plans, to comment on the outcome of existing management measures, to update descriptions of fisheries; to report on national sampling levels and data availability; and to consider measurement and estimation of misreporting and discards.

0.1 Working procedures

Much consideration was given to the problem of the increasing workload of the WG, both intersessionally and during the meeting itself. Several proposals were made by the Chair during the January 2006 Annual Meeting of Assessment Working Group Chairs (ICES- AMAWGC 2006). The stated intention to provide probabilistic assessments and forecasts did not materialise, as software is not yet available to do so, but an early data submission date was imposed with reasonable success and a substantial amount of intersessional work was carried out by WG members. The end of the meeting was reached with completed and checked text for all sections of the report, save the opening chapters (Sections 0-2) which were not completed until over a month afterwards. The information required by ACFM as the basis for advice was provided in good time for their October meeting.

As in the previous two years, the system of benchmark/update assessments was not closely followed by the WG. Ongoing developments in assessment methods and substantial revisions in stock perceptions following the inclusion of new data meant that pure update assessments were seldom appropriate. At the same time, the increasing workload reduces to almost zero the time available for the type of in-depth analysis that would be required for a benchmark analysis. Therefore, a pragmatic approach was taken: if intersessional work was done on an assessment, it became de facto a benchmark assessment, otherwise it was viewed as an update.

The order of the stock sections in this year’s report has been reversed from previous years, so that Nephrops are now discussed in Section 3 and cod in Section 14. This was done in an effort to encourage reviewers to devote as much time to stocks such as Nephrops, pout and sandeel (usually at the back of the report) as to stocks such as cod, haddock and whiting (usually at the front).

As last year, quality handbooks (stock annexes) for each stock are included in the main report as a series of appendices (appendix B3 – B14). This was done to avoid the problem of potentially useful stock-annex information being lost in the grey literature. In general these have not been modified this year, although there are exceptions.

0.2 State of the stocks

For Nephrops stocks, underwater TV surveys (where available) provided the best guide to state of stocks. The historical practice of basing numerical assessments on pseudo-ages was not followed. In TV-based Functional Units (FUs) Moray Firth (9), Firth of Forth (8), and Farn Deeps (6) abundance seems to be rising slightly. In Fladen Ground (7) abundance is fluctuating and currently towards mid to lower end of observed range. Other FUs were more difficult to assess (as there are no TV surveys) but in Noup (10), Norwegian Deeps (32), Botney Gut-Silver Pit (5) and Off Horn Reef (33) stocks seem fairly stable and no signs of overexploitation (LPUEs remain level and mean sizes are fairly constant). The harvest-rate

(14)

approach, based on F0.1, was proposed as the method for providing (and justifying) catch options where surveys were available. Other FUs and statistical rectangles outside of the main assessed areas were dealt with by status quo advice or mean of last three years landings. On this basis, the overall TAC for Sub-Area IV would be slightly reduced from last year. Status quo landings are advised for Division IIIa.

The directed fishery for Norway pout in Sub-area IV was closed during 2005 and most of 2006. Landings in 2005 (1.9 kt) were the lowest observed; these arose from experimental fishing and a limited bycatch. In-year survey-based monitoring in April 2006 led to the opening of the fishery with a TAC of around 90 kt, although less than 50% of this is likely to be taken. Estimated SSB for this stock in 2005 was well below Blim and fishing mortality was effectively zero. The size of the 2005 year-class was the largest since 1999, while the 2006 year-class was moderately abundant. The potential for a fishery in 2007 will be dependent on the survival and growth of these year-classes, along with the size of the 2007 year-class Landings in 2005 for sandeel in Sub-area IV (172 kt) remained at or near the same low level as in the preceding three years. Landings in 2006 have continued this trend, and following the implementation of a real-time management plan, the fishery was closed in July 2005.

Estimated SSB is close to its lowest observed level and is well below Blim. Fishing mortality has declined in recent years and is now below the long-term mean. Recruitment remains low.

In order to permit a fishery in 2007, the 2007 year-class would have to be substantially larger than recent year-classes.

Discrepancies between catch-at-age based analyses and survey-based analyses has prevented the WG from assessing the state of plaice in Division VIId. Landings have declined steadily since 2002 to 3500 tonnes, the lowest value since 1980.

Plaice landings in Division IIIa fell in 2005 to an historical low of 6905 tonnes. The available quota has never been restrictive for this stock. About 82% of the landings were taken in the Skagerrak. Although the assessment is uncertain, the WGs best estimates indicate that has fluctuated rapidly since 1996 and is currently relatively low (~ 0.85); and that SSB is increasing following recruitment of the large 2003 year-class.

As in the previous two meetings, the assessment of plaice in Subarea IV included modelled discard estimates for recent years. Landings and discards have both declined in recent years.

SSB remains at a relatively low level (between Blim and Bpa), while fishing mortality has declined (although it is still above the long-term mean). Recent year-class strength has been poor. Surveys suggest the 2005 year-class to be around the long-term average. On this basis, short-term forecasts at current fishing levels indicate a fall in landings in 2007 (to around 51 kt) and an increase in discards (to around 55 kt). For SSB to reach above Bpa by the start of 2008, landings in 2007 would need to be around 33 kt.

Landings for sole in Division VIId have fluctuated around a mean level for many years, and show no significant trends. The fishing mortality is estimated to be around Fpa The SSB has above Bpa (8000t) following improved recruitment in recent years, particularly of the year classes 1998 to 2000 and 2003. There is a tendency to underestimate F and overestimate SSB.

The reported landings for sole in Subarea IV in 2005 (16.4 kt) were at a similar level as in recent years. SSB has fluctuated around a moderate-to-low level for several years, although at status quo fishing mortality it is forecast to drop below Bpa during 2006. The short-term forecast at status quo F suggests a fall in landings (to around 12.5 kt in 2007) and a corresponding decline in SSB.

Reported landings for saithe in Subareas IV and VI and Division IIIa in 2005 were around the recent average (112 kt). The assessment was a standard update this year. Fishing mortality has now remained at or below 0.3 for six years (F ~ 0.26 in 2005) while SSB

(15)

continues a steady increase (288 kt in 2005). Recruitment is fluctuating about the mean level.

The TAC has been unrestrictive for four years. The short-term forecast as status quo F indicates landings of 109 kt in both 2006 and 2007 (2006 TAC ~ 136 kt), along with a slow decline in SSB (to 280 kt by the start if 2008).

Catches of whiting in Subarea IV and DivisionVIId continued to decline in 2005, and set a new historical low (21 800 kt). The whiting assessment is again quite uncertain. The same concerns as last year were raised about stock structure, but in the absence of improved information on stock distribution the WG decided to present the same approach as last year (in the full knowledge that this was rejected by ACFM). The final assessment indicates historically low estimates of recruitment (346 million), SSB (104 kt) and fishing mortality (0.25). Without good recruitment the stock is unlikely to recover. Short-term forecasts at status quo fishing mortality suggest falling landings (9100 tonnes in 2006, 8200 tonnes in 2007) and slowly increasing levels of both SSB and discards. The fact that the forecast landings for 2006 are less than half the permitted TAC raises concerns about the analysis.

This assessment must be considered in the light of industry reports that whiting are more abundant than for several years, particular off the north-east coast of England. The Scottish industry are also reporting good catches of whiting and are likely to take their quota in full, which doesn’t correspond to the low forecast landings for 2006.

The strong 1999 year-class again dominated the catches of haddock in Subarea IV and Division IIIa (57 300 kt), which were the lowest in the available time-series. The assessment (using the same procedure as last year) indicated a continued decline in SSB (from 298 kt in 2004 to 256 kt in 2005) as the 1999 year-class reduces in number. Fishing mortality has stabilised at or around 0.3 (it has now been in the range 0.25 to 0.35 for four years). The 2005 year-class (recruiting at age 0) is estimated to be quite abundant (35 000 million) and the largest since the 1999 year-class (now estimated to have been 114 000 million, a slight increase on the estimate in last year’s assessment). The WG considered the issue of appropriate inputs for the haddock forecast very carefully. In particular, the mean weights-at- age of the slow-growing 1999 and 2000 year-classes have now been modelled in a more realistic manner. The outcome at status quo fishing mortality in 2007 is landings of around 58 000 tonnes (compared with estimated landings in 2006 of 46 600 tonnes) and discards of 34 400 tonnes (compared with 20 500 tonnes in 2006). The increases in both catch components are due to the reasonably good 2005 year-class. While the increase in projected landings is good, the increase in discards is not and needs to be considered carefully if this year-class is to benefit the stock and the fishery for as long as possible.

The estimated yield (reported landings and discards) in 2005 for cod in Subarea IV and Divisions IIIa and VIId (40 300 kt) was low. A modified assessment has been used this year which is based on the combined survey series for the third quarter, and which uses an uncertainty estimation procedure. The assessment includes estimates of unaccounted removals, as for the last two years. Spawning-stock biomass remains low but stable (~ 35 kt).

Fishing mortality is now estimated to have declined since 2000 (median estimate for 2005 ~ 0.86). Recruitment of the 2000-2005 year-classes was poor. Indications from Q1 and Q3 surveys in 2006 are that the 2006 year-class is somewhat stronger. Results from a number of forecast scenarios covering different changes in TAC in 2007 indicate that SSB will continue to decline to a historic low. Only zero catch in 2007 will enable SSB to rise to Blim (70 kt) by the start of 2008.

0.3 Environmental and ecosystem considerations

The WG was asked to “consider existing knowledge on important environmental drivers for stock productivity and management and if such drivers are considered important for management advice, incorporate such knowledge into assessment and prediction, and important impacts of fisheries on the ecosystem.” This was addressed in each stock section,

(16)

where information was available to the WG. However, due to a lack of firm conclusions in the literature on causative mechanisms linking fish stocks and the environment, and poor predictability of ecosystems, few quantitative modifications were made to assessments or forecasts to account for environmental information. The exceptions were those stocks for which recent recruitment is clearly different (in some way) to historical recruitment, in which case the recent recruitment estimates only were used to generate recruitment forecasts. Apart from this, the report is limited to comments on potentially-important ecosystem impacts.

During the WG meeting a proposal was submitted by Martin Pastoors (Chair, ACFM) for consideration and comment, on possible restructuring of the WG to accommodate improved cognisance of ecosystem effects. Following a discussion in plenary, the WG reached a number of conclusions; these are summarised in Section 1.6 of this report.

0.4 Mixed-fisheries data collation and modeling

In previous years, a considerable amount of time has been spent during the WG meeting collating mixed-fisheries data, with little mixed-fisheries modelling. This year mixed- fisheries data was (to a certain extent) collated at the same time as single-species data, so the opportunity was taken during the WG meeting to explore the potential of the Fcube mixed- fisheries analysis system in a series of dry runs. These were not intended to be used as the basis for advice, but to indicate the strengths of the approach, and where it could be improved in the future. The provision of fisheries data in the appropriate fisheries aggregations remains the principal problem preventing the provision of mixed-fisheries management advice. The analyses are described in Section 15.

0.5 Management plan evaluations

A number of requests were received by ICES for the evaluation of management plans during 2006. Those regarding North Sea haddock, sandeel, Norway pout, plaice and sole were passed onto the WG for consideration, along with the standing request to evaluate the cod recovery plan. A substantial part of the WG meeting was devoted to this important issue, particularly during the second week, and the results and conclusions are provided in Section 16.

0.6 Data collation issues

The provision, exchange and raising of landings and discard data remains a serious problem for ICES assessment WGs, and is (in many respects) the most difficult issue that WGs have to deal with. Early submission of data this year ameliorated some of the difficulties, but introduced new ones as some data were not ready in time. Efforts are currently underway to address these concerns, and presentations were given to the WG on both InterCatch and FishFrame. ICES have insisted that InterCatch be used from 2007 onwards to collate WG assessment data. If this approach works then the provision of advice should run much more smoothly: however, there is a substantial training need to be met before this can become a reality. Further details on the WG discussion are given in Section 1.2.5.

(17)

1 General

1.1 Terms of reference

The Working Group on the Assessment of Demersal Stocks in the North Sea and Skagerrak [WGNSSK] (Chair: Coby Needle, UK) met in ICES Headquarters from 5-14 September 2006 to:

a) assess the status of and provide management options for the following stocks: 1) cod in Subarea IV and Division IIIaN (Skagerrak), and Division VIId, 2) haddock in Subarea IV and Division IIIa, 3) whiting and 4) plaice, both in Subarea IV, Division IIIa, and Division VIId, 5) saithe in Subarea IV, Subarea VIa, and Division IIIa, 6) sole in Subarea IV and Division VIId, for Norway pout and sandeel stocks in Subarea IV and Divisions IIIa and VIa, and 7) Nephrops stocks: Functional Units 3, 4, 5, 6, 7, 8, 9, 10, 32 and 33;

b) quantify the species and size composition of by-catches taken in the fisheries for Norway pout and sandeel in the North Sea and adjacent waters, and make this information available to the Working Group on Ecosystem Effects of Fishing Activities;

c) provide the data required to carry out multispecies assessments (quarterly catches and mean weights-at-age in the catch and stock for 2005 for all species in the multispecies model that are assessed by this Working Group).

WGNSSK, WGSSDS, WGHMM, WGMHSA, WGBFAS, WGNSDS, AFWG, HAWG, NWWG, WGNPBW and WGPAND will, in addition to the tasks listed by individual group in 2006:

1 ) based on input from e.g. WGRED and for the North Sea NORSEPP, consider existing knowledge on important environmental drivers for stock productivity and management and if such drivers are considered important for management advice incorporate such knowledge into assessment and prediction, and important impacts of fisheries on the ecosystem;

2 ) Evaluate existing management plans to the extent that they have not yet been evaluated. Develop options for management strategies including target reference points if management has not already agreed strategies or target reference points (or HCRs) and where it is considered relevant review limit reference points (and come forward with new ones where none exist) following the guidelines from SGMAS (2005, 2006), AGLTA (2005) and AMAWGC (2004, 2005, and 2006).

If mixed fisheries are considered important consider the consistence of options for target reference points and management strategies. If the WG is not in a position to perform this evaluation then identify the problems involved and suggest and initiate a process to perform the management evaluation;

3 ) where mixed catches are an important feature of the fisheries assess the influence of individual fleet activities on the stocks and the technical interactions;

4 ) update the description of fisheries exploiting the stocks, including major regulatory changes and their potential effects. Comment on the outcome of existing management measures including technical measures, TACs, effort control and management plans. The description of the fisheries should include an enumeration of the number, capacity and effort of vessels prosecuting the fishery by country;

5 ) where misreporting is considered significant provide qualitative and where possible quantitative information, for example from inspection schemes, on its distribution on fisheries and the methods used to obtain the information;

document the nature of the information and its influence on the assessment and predictions;

(18)

6 ) provide for each stock and fishery information on discards (its composition and distribution in time and space) and the method used to obtain it. Describe how it has been considered in the assessments;

7 ) report as prescribed by the Secretariat on a national basis an overview of the sampling of the basic assessment data for the stocks considered;

8 ) provide specific information on possible deficiencies in the 2006 assessments including, at least, any major inadequacies in the data on landings, effort or discards; any major inadequacies in research vessel surveys data, and any major difficulties in model formulation; including inadequacies in available software.

The consequences of these deficiencies for both the assessment of the status of the stocks and the projection should be clarified;

9 ) Further develop and implement the roadmap for medium and long term strategy of the group as developed by AMAWGC.

10 ) Working Group Chairs will set appropriate deadlines for submission of the basic assessment data. Data submitted after the deadline will be considered at a later meeting at the discretion of the WG Chair.

ToR a1 is addressed in Section 14, ToR a2 in Section 13, ToR a3 in Section 12, ToR a4 in Sections 6-8, ToR a5 in Section 11, ToR a6 in Sections 4, 5, 9 and 10, and ToR a7 in Section 3. Section 1.5 (Data for other Working Groups) provides the information requested in ToRs b and c. Of the additional ToRs to be addressed by all assessment WGs, ToR 1 was not covered due to a current lack of knowledge of causal relationships between the environment and marine fish stocks. For this reason, no quantitative modifications were made to assessments or forecasts to account for environmental information and the report is limited to comments on potentially-important ecosystem impacts. ToR 2 is covered in Section 16. ToRs 3 and 4 are addressed in each stock section (Sections 3 to 14), where information was available to the WG. More general data and analyses on mixed fisheries are given in Section 15.

Misreporting, discarding or other sources of unaccounted removals (ToRs 5 and 6) are considered in several stock sections. An overview of sampling rates and data availability for basic assessment data (ToR 7) is given in Section 1.2.4 and 1.2.5, while sampling for the purposes of mixed fisheries data collation is discussed in Section 15. Discussions regarding the quality of each assessment, in terms of data and modelling (ToR 8), are given at the end of each stock section. Considerations on the future strategy of the WG are summarised in Section 1.6. The results of an attempt to encourage adherence to deadlines for data submission are highlighted in Section 1.1.1.

1.1.1 Attempted changes in working practice

The workload of WGNSSK, in common with all assessment WGs, has been steadily increasing in recent years. For 2006 the WG has been asked to deal with ten generic ToRs, in addition to the traditional assessment and forecast requirements. It is clear that such a workload is unlikely to be addressed satisfactorily without substantial intersessional investment in time and resources. The Chair of WGNSSK made a comment to this effect in last year’s report (ICES-WGNSSK 2006), and raised the issue during the 2006 Annual Meeting of Assessment WG Chairs (ICES-AMAWGC 2006).

The Chair’s proposals on how the change the working practice of the WG are summarised in full in an appendix to the AMAWGC report. The key salient points are listed below, along with the actions that were taken this year to address them.

• In October 2006, ACFM decided that assessment WG Chairs would be allowed to set early submission dates for data. The intention here was to allow much more time than has previously been the case for intersessional exploratory assessments and mixed-fisheries analyses. The Chair accordingly set a date of June 30^th, 2006.

(19)

• As in previous years, a number of subgroups were set up which started work intersessionally and then acted as fora during the WG meeting for review and discussion of assessments and forecasts. These subgroups were organized deliberately so as to avoid the historical split into roundfish, flatfish, industrial and Nephrops divisions, with the intention that fresh approaches would emerge.

• However, subgroups were not used for reading and checking text. This was done in small plenary sessions, led by the Chair, which any WG member could attend (but none were forced to).

• Probabilistic assessments and forecasts, of the type discussed in ICES- AMAWGC (2006), were not used. The Methods WG (ICES-WGMG 2006) concluded that the approach taken was useful as a sensitivity analysis, but could not be used as the basis of probabilistic assessments.

This attempt to change working practice has been moderately successful. Most data were provided in time, although fishery-based data still caused great difficulties and some countries with hitherto good records (such as Scotland) encountered database difficulties that slowed submission. Preliminary assessments for nearly all stocks were presented during the first two days of the WG meeting, and final assessments were in most cases completed by the weekend, thus allowing due attention to be paid to forecasts. The new approach also allowed for much more analysis on the issues of mixed fisheries modeling (Section 15) and management plan evaluations (Section 16). Subgroups worked successfully, and the plenary text read-throughs meant that the Chair had a very good overview of all the assessment issues raised. On the other hand, this reduced to almost nil the time available to the Chair for writing text and organising the final report, which was finished well after the end of the meeting as a consequence. Further discussion on the possible future structure of the WG is given in Section 1.6.

It should be noted that there was no French participation in this year’s WG meeting.

Although a working paper was submitted for the plaice VIId assessment (see Section 6), there was no input from France to discussions of fisheries and stock perceptions. This hindered the WG, particularly for stocks such as Channel plaice and sole, saithe, and North Sea whiting which include important French fisheries. This development is a real concern for the WG and for ICES, and needs to be addressed carefully for next year.

1.2 Data sources and sampling levels

1.2.1 Roundfish and flat-fish stocks

The data used in assessments for stocks of roundfish (cod, haddock, whiting, saithe) and flatfish (plaice, sole) are based on:

• total reported landings by market size categories;

• sampling programmes for weight, length, age, and sometimes maturity, by market size categories;

• observer sampling programmes for discards;

• effort data from logbooks, and catch-per-unit effort (CPUE) or landings-per-unit effort (LPUE) data from associated fleet landings;

• research-vessel survey indices by age; and

• data on natural mortality from multispecies analyses.

1.2.1.1 Landings, age compositions, weights-at-age, maturity

In a number of cases, management areas do not correspond exactly with the areas for which the assessments are carried out. If the management areas are larger, landings cannot always be obtained for the assessment areas separately. In these cases landings have to be estimated by the Working Group (WG) from external information.

(20)

For most stocks, the WG estimates of total landings deviate from official figures. The discrepancies are shown in the landings tables in the relevant stock section, under the heading unallocated landings. These unallocated landings will in most cases include discrepancies that are due to differences in calculation procedures. For instance, in some cases national conversion factors from gutted to live weights have been changed in the official statistics, but not in the WG database. The differences introduced by conversion factors, and the difference between sums-of-products (SOP) of landed numbers and estimated mean weights on the one hand, and nominal landings on the other, may arise through inadequate sampling or data reporting, and are minor in most cases. SOP corrections are applied in some cases for the flatfish stocks, where deemed necessary, and are a standard procedure for all roundfish stocks.

In a number of cases, uncertainties in the landing data can seriously affect the quality of the assessments and catch forecasts. In some cases, the WG estimates of the landings include specific corrections for misreported or unreported landings. These are discussed in the relevant Stock Annex sections of the Quality Control Handbook (included as an appendix to this report). There are signals that unallocated removals of various kinds occur in other stocks, especially in the stocks of valuable species: these removals may be due to fisheries (unrecorded discards, misreporting, or non-reporting) or to ecosystem changes. However, by their nature these could not be verified or quantified. Continued concerns about the quality of North Sea cod landings data in particular have been addressed in this year’s report (Section 14) by the use of an assessment method which estimates the magnitude of unallocated removals via research-vessel survey information.

Historical time-series (aggregated at the fleet level) of age compositions, weights-at-age, and length-at-age are archived, maintained and collated in databases at national institutes.

Roundfish data (cod, haddock, whiting, and saithe) are collated in Aberdeen (FRS). North Sea plaice and sole are maintained in IJmuiden (RIVO), VIId plaice in Port-en-Bessin (IFREMER), VIId sole in Oostende (DVZ), and IIIa plaice in Charlottenlund (DIFRES). Any revisions that have been made to these data are indicated in the relevant stock sections.

The countries that are responsible for the major proportions of the total landings for each stock generally provide the age composition data for those stocks. For the years up to and including 2001, each country was obliged to sample only national vessels. This meant that foreign vessels landing abroad were not sampled. The sampling procedure was changed to address this problem, and from 2002 onwards each country has been required to sample (where possible) the landings of all fleet components landing in their country (EU regulation 1639/2001).

Mean weights-at-age are either derived from observations of catch weights-at-age (for flatfish and industrial species), or from fixed weight-length relationships applied to observations of length distributions from catches (for roundfish). In most stocks the annual mean weights-at- age in the stock are set equal to the mean weights-at-age in the catch, due to lack of fishery- independent information on weights. Exceptions are the North Sea and eastern English Channel plaice and sole stocks for which the weight-at-age in the stock is set equal to the weight-at-age in the first quarter (plaice) or second quarter (sole). For all stocks, the mean weights-at-age in the catch of the youngest age groups may not accurately represent the mean weights-at-age in the stock due to fisheries selecting for larger fish.

Estimates of the proportion mature-at-age (maturity ogives) are based on historical biological information and are kept constant over the whole time period of the assessment. For a number of stocks a knife-edged maturity ogive has been assumed. Observations on maturity-at-age (from research-vessel surveys, for example) indicate that the age of maturation can change over time. The assumption of constant maturity ogives may introduce bias in estimated spawning-stock biomass (SSB), especially when exceptionally large or small year classes enter the spawning stock.

(21)

1.2.1.2 Discards

Estimates of discards are used in the assessments for cod, haddock, whiting and plaice in the North Sea. All the discard data for other species that was made available to the WG has been presented in the report (see the relevant stock sections), although they appear to be based on sampling that is too sparse to permit their inclusion yet. There is a continuing discrepancy between the observer sampling required by European legislation, and the data made available to ICES WGs, and this needs to be addressed as a matter of urgency.

The use of discard estimates in assessments is thought to reduce bias, give more realistic estimates of fishing mortality, and lead to more representative inputs for mixed fisheries analyses. However, discard estimates can be noisy and increase the variability of the assessment. Furthermore, for many of the stocks it is unclear whether the available discard estimates form a representative sample of discarding practice in the fisheries.

For cod, haddock and whiting, total annual international discard estimates by age group were derived largely by extrapolation from the Scottish discard sampling programme. Data from other sampling programmes were made available for this process, but not in a form that could be used in the roundfish discard collation procedure. Discard estimates for plaice in the North Sea were obtained by a combination of observations from the Dutch and English beam-trawl fisheries for recent years, and reconstructions based on observed growth for earlier years (see Section 8).

Problems with data collation procedures are discussed in Section 1.2.5.

1.2.1.3 Natural mortality

Natural mortality cannot readily be distinguished from fishing mortality by analyses of catch-at-age and research-vessel survey data. Therefore, unless stock analysis is conducted on the basis of total mortality (as is the case with the SURBA model, Section 1.3.3), natural mortality must be estimated separately from the assessment procedure. The estimates of natural mortality for cod, haddock and whiting are based on historical estimates of multispecies predation rates (ICES-MAWG 1989) and, unless specified otherwise, are kept constant over the whole time period of the assessment. In the plaice and sole stocks, natural mortality is assumed to be 0.1 for all age groups (with an exception for sole to account for the cold winter of 1963). The natural mortality of saithe is assumed to be 0.2 for all age groups, and at 0.4 per quarter for all age groups of Norway pout (although this is discussed further in Section 5). For sandeel, the natural mortalities used are derived from multispecies considerations, although they are not exactly the same (see the sandeel Stock Annex Q4).

1.2.1.4 Commercial fleet and research vessel data

All available time-series of CPUE and effort data from commercial fleets and research-vessel surveys have been presented in this year’s report, and a subset of these data have been used to calibrate catch-at-age-based assessments and short-term forecasts (see Table 1.3.2). For most stocks, survey-based assessments have also been presented as exploratory analyses.

The validity of many of the commercial tuning fleets as indicators of stock size and fishing mortality in recent years has become more uncertain, since the enforcement of national quota, ITQs, and technical measures is known to have led to changes in fishing patterns (and in some cases to possible misreporting and discarding). For this reason, commercial CPUE data has been excluded from the assessments of a number of stocks. Such data has been retained in assessments only in cases where no survey data are available, or where commercial CPUE series provide reliable information that cannot be obtained elsewhere. At the time of year when the meeting took place, survey indices from the Dutch beam trawl survey, the IBTS Q3 survey and the English Q3 groundfish survey were not available. The latter was due to be ready for several stocks (beginning with North Sea cod) by the end of September 2005.

(22)

Figure 1.2.1 shows the roundfish sampling areas covered by the IBTS Q1 and Q3 surveys.

1.2.2 Norway pout and sandeel

The data used in the assessment for Norway pout and sandeel stocks are based on:

• total landings;

• samples of landings for species composition, weight, length, age, and sometimes maturity. Samples of industrial landings are used for an exact species composition of by-catch species and to get the percentage of target-species;

• fleet data: effort data from logbooks and CPUE data from associated fleet landings;

• survey data: survey indices by age for Norway pout;

• data on sandeel natural mortality from the MSVPA.

1.2.2.1 Landings, age compositions, weights-at-age, maturity

The sampling of Norway pout and sandeel landings are described in detail in the relevant Quality Control Handbooks (see Annexes Q4 and Q5). The applied sampling systems vary between countries.

In Norway, the sampling system since 1993 has been based on catch samples from three market categories: E02 (mainly sandeel), D13 (blue whiting, if not sandeel and catch taken west of 0°E), and D12 (Norway pout, if not sandeel and catch taken east of 0°E). The samples are raised to total landings on the basis of sales slip information on landed categories. Effort is estimated from the total number of trips and an estimate of average days-at-sea per trip.

In Denmark, the catch estimates are based on sales slip information, logbook data, species composition from inspectors, and biological data, including age-length keys from independent biological sampling. Total landings are estimated per statistical rectangle based on total catch estimates from sales slip and logbook data, together with biological and species composition data. Historical time-series of market sampling data for sandeel and Norway pout are kept and maintained in Charlottenlund (DIFRES). Any revisions in the catch- and weight-at-age data are indicated in the relevant stock sections.

In the assessment of Norway pout the weights-at-age in the stock are kept constant over the whole period of assessment. Samples from the landings, however, suggest high variability both between years and between seasons. One of the problems of using mean catch weights is that the 0-group is not fully recruited in the third quarter, giving an overestimate of weight-at- age in the stock for this age group. More knowledge is required before variable weight-at-age in the catches can fully be taken into account in the assessment. For sandeel, the weights-at- age in the catches in the first half-year are used as estimation for weights-at-age in the stock.

The maturity ogives for Norway pout and sandeel are kept constant over the whole period of assessment (although see discussion of maturity estimates for Norway pout in Section 5).

Natural mortality estimates are based on historical information and kept constant over the whole time period of the assessment. Values are given in the relevant stock sections.

For Norway pout, time-series of CPUE and effort data from Danish and Norwegian commercial fleets and data from research vessels are available. The research vessel data include the IBTS Q1 and Q3 series, and the Scottish and English Q3 series.

(23)

For sandeel, only data from the Danish and Norwegian commercial fleets are available.

Indices from research-vessel surveys are in development for sandeel, and are described in Section 4.9.

1.2.3 Nephrops

1.2.3.1 Landings, length frequencies

Length and sex compositions of Nephrops landings are estimated from either port or onboard sampling. Length data are applied to all catches and raised to total international landings.

Rates of discarding by length class are estimated by on-board sampling or shore based sampling of total catch, and extrapolated to all other fleets.

The differences in catchability between sexes have lead to the two sexes being assessed separately. And hence removals are raised separately for each sex. Trawl and creel fisheries are sampled separately.

In the absence of routine methods of direct age determination in Nephrops, age compositions of removals were inferred from length compositions by means of ‘slicing’. This procedure, introduced at the 1991 Nephrops WG, uses von Bertalanffy growth parameters to determine length boundaries between age classes. All animals in length classes between boundaries are assigned deterministically to the same age class. The method is implemented in the L2AGE programme which automatically generates the VPA input files. The programme was modified in 1992 to accommodate the two-stage growth pattern of female Nephrops and again in 2001 to separate ‘true’ as opposed to ‘nominal’ age classes). The age classes are ‘true’ to the extent that the first slicing boundary, i.e. lower length boundary for ‘age’ 0, is the length-at-age zero rather than the lowest length in the data. This ensures comparability of ‘age’ classes across stocks. The output from this procedure was used as part of the analyses to generate appropriate harvest rates, rather than in assessments per se.

1.2.3.2 Discards

Discard data are available for a number of Nephrops stocks, generally collected on a quarterly basis by Functional Unit. Landings and discards at length are combined (assuming a discard survival rate of 0-25%, depending on the stock) to removals.

A natural mortality rate of 0.3 is assumed for all age or length classes and years for males and immature females, with a value of 0.2 for mature females. The lower value for mature females reflects the reduced burrow emergence while bearing eggs, and hence an assumed reduction in predation.

Landings at age and effort data for various national Nephrops trawl fleets are used to generate CPUE or LPUE indices. Catch at age are estimated from raising length sampling of discards and landings to officially recorded landings, and slicing into ages (knife edge slicing using growth parameters). CPUE is estimated using officially recorded effort (hours fished) although there are concerns over the accuracy of landings and effort for some stocks. There is no account taken of any technological creep in the indices.

Underwater TV survey: The burrowing nature of Nephrops, and variable emergence rates mean that trawl catch rates may bear little resemblance to population abundance. An underwater TV survey has been developed, estimating Nephrops population abundance form burrow density raised to stock area. A random stratified sampling design is used, on the basis

(24)

of sediment strata and a regular grid. The survey provides a total abundance estimate, and is not age or length structured.

1.2.4 Sampling levels and procedures

Methods of data collection and processing vary between countries and stocks. The sampling procedures applied in the various countries to the various stocks until 2002 were described in detail in the report of the WGNSSK meeting in 1998 (ICES-WGNSSK 1998). Since 2002 an EU regulation (1639/2001) has been in place which has altered market sampling procedures.

Firstly, each country is obliged to sample all fleet segments, including foreign vessels, landing in their country. Secondly, a minimum number of market samples per tonnes of landing are required. The national market sampling programmes have been adjusted accordingly.

Table 1.2.1 gives an overview of the sampling levels in 2005 for each stock, for both landed and discarded components of catch.

1.2.5 Data collation (Intercatch, FishFrame) and current problems

One of the key difficulties for the WG is the acquisition and collation of data on which to base assessments, forecasts and other analyses. The collation procedures for single-stock analyses have become increasingly antiquated in recent years, a trend worsened by a marked difference in approach between different subtypes of demersal species (roundfish, flatfish, Nephrops and industrial fish all have different data collation procedures). The problem has been exacerbated in the last two years by increased calls for mixed-fisheries (i.e. fleet-based) landings and discards data. Some of these data are simply not available. Others are not made available to the WG for one reason or another, or they may be available but in the wrong format. Lack of resources in staff time hinders data collation in many cases.

The EU Data Collection Regulation (DCR) is intended to rectify these problems. In some cases it seems to have been only partially successful. Fisheries data, particularly discard data, which countries are paid to collect and provide to ICES are not made available to the relevant WGs. Countries which do provide data on discards are highlighted as discarding fish by the EU, leading to increased legislation and an understandable reluctance to participate in observer sampling schemes (seen as self-incriminatory in some quarters). The EU-STECF working groups appear to be more able to acquire fisheries data, perhaps because such data must be destroyed 20 days after the end of the relevant meeting.

In the revised MoU between ICES and the EU, a new clause stipulates that any data collected under the DCR must be made available to ICES. It is hoped that this will alleviate some of the aforementioned difficulties. To check progress of this enforced rigour, and to try and deliver quality assurance, the EU have requested that ICES completes reformatted data availability tables for each stock. These are similar to those in Table 1.2.2 but include much more detail on which countries are supplying which data, and for what purpose. ICES presented these new tables during the 2006 WG meeting and requested that these be completed within one week of the close of the meeting. This proved to be an unworkable request, due to considerable time constraints, but the WG agrees that these are worthwhile summaries of data availability and that they will be completed at the 2007 meeting.

Several initiatives are underway to try and address these problems. Two of these were presented at the meeting:

1 ) Intercatch. Henrik Kjems-Nielsen (ICES) gave a presentation of recent developments in the Intercatch system, in which he emphasised that ICES expects WGs to use the system from 2007 onwards. He clarified what the system is intended to do: namely, to act as a database tool in which stock coordinators can collate national datasets in order to generate assessment files (single-stock and mixed-fisheries) for WGs. The system will not hold raw

(25)

national data – the intention is that this will be worked up to the national level using whichever system national data collators wish to use (e.g. FishFrame, see below) prior to submission to InterCatch. The WG raised several issues which were addressed which varying degrees of success:

1.1 ) It was not clear to the WG who was to chase up national data providers.

The ICES response was that this is still very much the responsibility of the stock coordinators – in fact, using Intercatch actually increases the reliance of the whole system on the coordinators. Their training is therefore of pivotal importance, and needs to be addressed by ICES.

1.2 ) Fleet and fishery definitions are still in development, which hinders the creation of fishery-based datasets. The ICES view on this is that fishery definitions can (in the first instance) be very rudimentary, thus allowing the system to operate until such time as improved fishery definitions are agreed.

At the moment, the implications for mixed-fisheries analyses of the Intercatch system are not yet clear.

1.3 ) The question of data ownership was raised. As mentioned above, STECF collates data under an agreement that said data will be destroyed 20 days after the end of the relevant meeting. Some countries may object to submission of data to a permanent ICES database. This question was not resolved by the end of the WG meeting.

2 ) FishFrame. A summary of developments in FishFrame was presented by Henrik Degel and Teunis Jansen of DIFRES (Charlottenlund). This system can be used in several different ways, but one approach is for national data collators to use it as a tool for working up sampled data to national fleet level (allowing for the correct fill-ins for missing age-length keys, etc). They emphasised that, unlike Intercatch, use of FishFrame is entirely voluntary. Clearly testing in the North Sea context is required (FishFrame was largely developed for use in the Baltic), but equally clearly it has the potential to replace several of our current aging data collation systems. As for Intercatch, training and data ownership are key issues to be addressed. A series of workshops have been given and more are planned, while data access issues may be circumvented by downloading local copies of the FishFrame system. FishFrame has the additional burden of requiring funding (as DIFRES are unable to continue maintaining and developing it for ever).

Applications are being prepared for EU funding.

1.2.6 Developments and changes to IBTS series collation

IBTS survey indices have changed slightly since last year due to data and code updates. The extent of these changes is dependent on species. The following points explain why.

1 ) For the years 2003 to 2005 ICES received revised data from two countries. A number of problems were found in these datasets. Some of them were related to misunderstanding of the exchange format and some were extractions problems from the countries’ database to the DATRAS exchange format. This affected both the CPUEs and the ALKs.

2 ) ICES are at the moment implementing a bootstrap procedure for calculating the variance of the indices. The performance of the “hole filling” in the ALK was running very slowly and in order to bootstrap ICES had to improve this part of the calculations of the indices. When going through the “hole filling” algorithm there were found to be two problems in the old code:

2.1 ) There was a small error that might effect some of the last ages on the second decimal.

2.2 ) The procedure for filling the gap between the minimum length and the first observation was changed. A few fish have therefore moved from first to second year class (e.g. from age 1 to 2 in quarter 1).

3 ) The data type has changed in some of the procedures from float to decimals. This can give a difference of 0.001 when the indices are summed over all ages in a few cases.

(26)

4 ) In 1992 and 1983 there were missing substitutions keys for Norway pout and haddock in area 8 and 9.

The reliability of the separate Scottish and English groundfish Q3 surveys (which form part of the IBTS Q3 index) as cod abundance indices was questioned in last year’s report (ICES- WGNSSK 2005), and further in Section 14 of this year’s report. This year’s cod assessment uses the IBTS Q3 data instead of the separate Scottish and English series; however the cost of this is the removal from analysis of autumn survey data in the current year. During discussions at this year’s meeting, ICES indicated that they would be able in future to provide IBTS Q3 data in time for the autumn ACFM meeting. WGNSSK would fully support this timetable and recommends that it be implemented.

1.3 Methods and software

1.3.1 Update and benchmark assessments

ACFM has requested that assessment WGs work to an agreed schedule of update and benchmark assessments. After experiencing problems in 2004 trying to accommodate a strict split between update and benchmark assessments, the WG has taken a different approach during 2005 and 2006. The large number of stocks and ToRs that the WG is asked to address means that the scope for in-depth analysis during the meeting itself is very limited, so that the range of approaches that would be expected in a full benchmark cannot be fulfilled. At the same time, stocks and fisheries in the areas covered by the WG are in such rapid flux that a simple update assessment is seldom appropriate. An update is also inappropriate if the assessment is to be reviewed externally. Therefore the majority of the assessments produced by the WG this year are neither update nor benchmark assessments, but somewhere in between. The range of analyses available in each stock section reflects the amount of work that could be done intersessionally on each stock rather than strict adherence to a predefined timetable. In other words: if intersessional work is done on a stock assessment, then that assessment is treated as a de facto benchmark; otherwise it is an update.

1.3.2 Quality control handbooks

Stock annexes (included in this report as Annexes Q3 to Q14) have not in general been updated this year (although there are exceptions). The new format of the first part of each stock section (introduced for the first time in ICES-WGNSSK 2005) has meant that some information (on ecosystem aspects and fisheries, principally) which previously would have been kept within the stock annexes has now been moved to the stock sections. Due to time constraints, most of these stock annexes have not been modified accordingly, so there may be some repetition. As before, the WG intends to undertake a full revision of stock annexes intersessionally.

1.3.3 Assessment methods

Table 1.3.1 lists the biological basis of the stock assessments undertaken by this Working Group. Table 1.3.2 gives an overview of model settings for these assessments.

XSA and SXSA

Extended Survivors’ Analysis (XSA; Darby and Flatman 1994) has been used for catch-at-age analysis for most stocks, although it has not been selected as the final assessment in all cases.

Three implementations were used. The version (FLXSA) incorporated in the FLR package (FLR Team 2006) was used in many cases to perform exploratory analyses. To date this implementation cannot produce standard output for tuning diagnostics, so version 3.1 of the Lowestoft VPA package was used for generating final runs. Seasonal XSA (Skagen 1993,

ACFM3506.pdf (19.72Mb)

ICES A

C

F

M

ACFM:35

R EPORT OF THE W ORKING G ROUP ON THE

A SSESSMENT OF D EMERSAL S TOCKS IN THE

N ORTH S EA AND S KAGERRAK (WGNSSK)

5–14 S EPTEMBER 2006

ICES H EADQUARTERS

Contents

R EPORT OF THE W ^ORKING G ROUP ON THE

A SSESSMENT OF D ^EMERSAL S TOCKS IN THE

N ^ORTH S ^{EA AND} S ^KAGERRAK (WGNSSK)